Pharmacological manipulation of the purinergic P2X7
receptor: effects on tumor and host cells
Dissertation
Submitted to the Department of Chemistry,
Faculty of Mathematics, Informatics and Natural Sciences, University of Hamburg
for the degree of
Doctor of Natural Sciences (Dr. rer. nat.)
by
Sana Javed
From Faisalabad, PakistanSupervisor:
Prof. Dr. med. Friedrich Haag
University Medical Center Hamburg-Eppendorf Department of Diagnostic
Institute of Immunology, Martinistrasse 52,
20246 Hamburg, Germany
Co-supervisor:
Prof. Dr. rer.nat Peter Heisig University of Hamburg Department of Chemistry
Institute of Biochemistry and Molecular Biology Bundesstrasse 45
20146 Hamburg, Germany
The present work was carried out in the period from October 2014 to December 2019 at the Institute for Immunology at the University Clinic Hamburg-Eppendorf in the working group of Prof. Dr. Friedrich Haag.
Table of Contents Zusammenfassung ... 6 Abstract ... 8 Abbreviation ... 10 1 Introduction ... 13 1.1. Purinergic receptors ... 13 1.1.1. P2X7 ion channel ... 13 1.2. Purinergic signaling ... 14 1.3. P2X7 in cancer ... 16
1.4. P2X7 blocking and potentiating nanobodies ... 17
1.5. Immunogenic cell death (ICD) ... 17
1.5.1. Unfolded Protein response ... 20
1.5.2. Autophagy ... 21
2 Aims of the project ... 22
3 Results ... 23
3.1. In-vitro results ... 23
3.1.1. P2X7 expression in Yac-1 cells ... 23
3.1.2. P2X7-mediated CD62L shedding, PS flashing and DAPI uptake ... 24
3.1.3. ATP-mediated cytotoxicity is blocked by P2X7 antagonist 13A7 ... 25
3.1.4. Gating of P2X7 enhances the uptake and cytotoxicity of doxorubicin in Yac-1 lymphoma cells ... 26
3.1.5. Doxorubicin cytotoxicity in Yac-1 cells is independent of P2X7 activation. ... 27
3.1.6. Gating of P2X7 enhances uptake of doxorubicin at early timepoints. . 27
3.1.7. ATP-induced cell death is rapid, but death due to ATP/DOX synergism is a late event ... 28
3.1.8. Transient activation of P2X7 marks cells for doxorubicin-induced death 29 3.1.9. Synergistic effects of P2X7 signaling on doxorubicin toxicity are dependent on the P2X7 isoform ... 31
3.1.10. P2X7 (k)-expressing A20 B lymphoma cells also show enhanced doxorubicin cytotoxicity in the presence of ATP ... 33
3.1.11. P2X7 and doxorubicin co-operate in the induction of phagocytosis-promoting signals ... 35
3.1.12. P2X7 also enhances the cytotoxicity of Bortezomib, a known inducer of the unfolded protein response ... 38
3.1.13. P2X7 and doxorubicin synergize to activate Caspase-3 ... 39
3.1.15. P2X7-mediated CRT translocation, PS exposure, and pore formation
are regulated by the PERK pathway ... 40
3.1.16. Inhibition of PERK blocks the synergistic effect of P2X7 on doxorubicin-mediated cytotoxicity ... 42
3.2. In-vivo results ... 43
3.2.1. Expression of P2X7 inhibits tumor growth. ... 43
3.2.2. Expression of functional P2X7R slows 4T1 tumor growth in-vivo ... 45
3.2.3. P2X7 enhancing nanobody slows down the tumor growth. ... 48
3.2.4. Subsets of tumor-infiltrating lymphocytes... 49
3.2.5. Phenotypic analysis of tumor-infiltrating lymphocytes ... 50
3.2.6. Subsets of lymphocytes in the spleen ... 54
3.2.7. Phenotypic analysis of T cells in the spleen ... 56
3.2.8. Increased co-expression of CD39 and CD73 on tumor infiltrating Lymphocytes (TIL) ... 57
4 Discussion ... 59
4.1. Synergistic cytotoxic effects ... 59
4.2. Possible mechanisms of synergistic cell death ... 61
4.2.1. Increased Drug uptake ... 61
4.2.2. ER stress and immunogenic cell death ... 61
4.3. P2X7 expression on tumor cells ... 64
5 Materials and Methods ... 68
5.1. Materials ... 68
5.2. Methods ... 76
5.2.1. Methods in Molecular Biology ... 76
5.2.2. Methods in cell biology ... 81
5.2.3. In-vitro assays ... 85
5.2.4. Methods in animal experiments ... 87
5.3. Statistical analysis ... 89
6 References:... 90
7 Appendix ... 99
7.1. List of Figures ... 99
7.2. List of tables ... 101
7.3. Safety and Disposal ... 102
7.4. Curriculum vitae ... 106
7.5. Acknowledgement ... 108
Zusammenfassung
Der purinerge P2X7-Rezeptor ist ein ATP-gesteuerter Kationenkanal, der sowohl von hämatopoetischen als auch von vielen Tumorzellen exprimiert wird. Die Expression von P2X7 bietet einen Wachstumsvorteil für Tumorzellen, indem der mitochondriale Calciumgehalt erhöht und der Zellstoffwechsel stimuliert wird. Andererseits kann eine starke Stimulierung von P2X7 zum Tod von Tumorzellen führen, und die Expression von P2X7 durch Wirtsimmunzellen ist für den Tumor schädlich, indem sie die Antitumor-Immunantwort verstärkt. Über die Rolle von P2X7 in der Chemotherapie ist wenig bekannt. Ich untersuchte daher die Auswirkungen der P2X7-Stimulation in Gegenwart von Chemotherapeutika die einen immunogenen Zelltod auslösen, wie z.B. doxorubicin (DOX) und bortezomib (BTZ).
Zur Untersuchung der In-vitro-Toxizität wurden Yac-1-Lymphomzellen, die P2X7 endogen exprimieren, sowie weitere mit P2X7-Varianten stabil transduzierte Zelllinien (A20 B-Lymphom, 4T1 Mammakarzinom) verwendet. Niedrige Dosen von extrazellulärem ATP (eATP), die selbst eine geringe Toxizität verursachten, erhöhten die Empfindlichkeit gegenüber DOX um das 7- bis 8-fache. Eine einstündige vorübergehende Exposition gegenüber eATP und DOX reichte aus, um den Zelltod nach einer Kultur über Nacht auszulösen. Mechanistisch erhöhte das Gating von P2X7 zwar die anfängliche Aufnahme von DOX in Zellen. Der verstärkte Zelltod konnte jedoch durch Blockade von PERK, einer zentralen Kinase des ER-Stress Signalweges abgeschwächt werden, was darauf hindeutete, dass der Synergismus eher auf einer Wechselwirkung nachgeschalteter Signalwege beruhte. Die Aktivierung von P2X7 durch ATP führte zur Phosphorylierung des PERK-Substrats eIF2a und zur Translokation von Calreticulin an die Zelloberfläche, zwei Kennzeichen des immunogenen Zelltodes.
Als Tumormodelle für in vivo Untesuchungen wurden Yac-1 T-Lymphom Zellen nach i.v. Injektion in SCID Mäuse sowie 4T1 Mammakarzinom-Zellen (stabil retroviral transduziert mit funktionalem oder nicht-funktionalem P2X7) nach lokaler Injektion in das Brustfettgewebe von Balb/C Mäusen verwendet. Dabei wurde die Tumorgröße im Yac-1 Modell durch Bildgebung, im 4T1 Modell durch einen Messchieber gemessen. In beiden Modellen verlangsamte die Expression von funktionalem P2X7 durch Tumorzellen das Tumorwachstum
Meine Ergebnisse zeigen, dass P2X7 auf Tumorzellen zum Erfolg der Chemotherapie beitragen kann, und legen nahe, dass die Modulation der P2X7 Aktivität eine mögliche therapeutische Strategie zur Erhöhung der Zytotoxizität von Chemotherapeutika darstellt.
Abstract
The purinergic P2X7 receptor is an ATP-gated cation channel expressed by various immune and tumor cells. Expression of P2X7 by tumor cells can favor tumor growth by raising mitochondrial calcium levels and stimulating the cell metabolism. On the other hand, strong stimulation of P2X7 can cause cell death. Expression of P2X7 by antigen presenting cells (APCs) results in maturation of dendritic cells and thus is detrimental to the tumor by enhancing the anti-tumor immune response. Little is known about the role of P2X7 in the context of chemotherapy. Therefore, I studied the effects of P2X7 stimulation in the presence of chemotherapeutic agents that induce immunogenic cell death (i.e., doxorubicin (DOX) and bortezomib (BTZ).
For in-vitro cytotoxic studies I used Yac-1 lymphoma cells that express P2X7 endogenously and also other cell lines (A20 B lymphoma cell line, 4T1 mammary carcinoma cell line) that were transduced with P2X7 variants. Low doses (100 to 200 µM) of extracellular ATP (eATP) that were not toxic by their own synergistically enhanced the sensitivity to DOX and BTZ. Transient exposure to eATP and DOX for one hour was sufficient to induce cell death after overnight culture. Mechanistically, gating of P2X7 augmented the initial uptake of DOX into cells. However, a decrease in cell death was observed when an important pathway of unfolded protein response, PKR ER kinase (PERK) was blocked by using PERK inhibitor, suggesting that the synergism was rather due to an interaction of downstream signaling pathways. I observed that ATP mediated P2X7 activations resulted in the phosphorylation of eIF2a as well as calreticulin (CRT) translocation to the cell surface, which are important hallmarks of immunogenic cell death.
For in-vivo studies, tumor graft models were prepared using Yac-1 lymphoma cells (injected i.v. in SCID mice) and 4T1 breast cancer cell line (stably transduced with functional or nonfunctional P2X7 by using retroviral vector) injected in mammary fat pad of Balb/C mice. In Yac-1 injected mice, tumor size was measured using optical image system while in the 4T1 injected mice, tumor volume was assessed by using caliper. In both animal models expression of P2X7 receptor by tumor cells slowed down the tumor growth.
My results showed that P2X7 activation synergistically enhance the cytotoxic effects of chemotherapy in vitro. Expression of P2X7 by tumor cells helped in tumor cell killing, both in-vivo and in-vitro. Therefore, P2X7 receptor activation on tumor cells could be considered as a combination therapy for better therapeutic outcomes.
Abbreviation
°C degrees centigrade
ADP Adenosine diphosphate AMP Adenosine monophosphate
APC Allophycocyanin
APC Antigen presenting cells ART2 ADP-ribosyl transferase 2 ATF4 Activating transcription factor 4 ATF6 Activating transcription factor 6 ATP Adenosine triphosphate
BIP Binding immunoglobulin protein BSA Bovine serum albumin
BTZ Bortezomib
cAMP Cyclic adenosine monophosphate CD Cluster of differentiation
cDNA Cyclic deoxyribonucleic acid
CHOP CAAT/enhancer-binding protein (C/EBP) homologous protein
CRT Calreticulin
DAMP Damage associated molecular pattern
DC Dendritic cells
DOX Doxorubicin
DMEM Dulbecco’s modified eagle medium DNA Deoxyribonucleic acid
dNTPs Deoxyribonucleotide triphosphate EDTA Ethylenediaminetetraacetic acid eGFP Enhanced green fluorescent protein eIF2a Eukaryotic initiation factor 2 alpha ER Endoplasmic Reticulum
ELISA Enzyme linked immunosorbent assay FACS Fluorescence activated cell sorting FCS Fetal calf serum
Foxp3 Forkhead box protein 3 FSC Forward scatter
HEPES 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid
g gram
HLE Half-life extended i.p. intraperitoneal IFN-g Interferon g
Ig Immunoglobulin
IL Interleukin
IL-1b Interleukin 1b
IRE1 Inositol-requiring enzyme-1
i.v. Intravenous Da Daltons l Liter Lb Luria broth M Molar mg milligram
MHC Major histocompatibility complex
min Minute
ml milliliter
mM millimolar
mP2X7 Mouse P2X7
mRNA Messenger ribonucleic acid
NAD Nicotinamide adenine dinucleotide
NTPDase Ectonucleoside triphosphate diphosphohydrolase NPP Nucleotide pyrophosphatase phosphodiesterase
ns Non-significant
PBS Phosphate buffered saline PCR Polymerase chain reaction PERK PRK like ER kinase
RNA ribonucleic acid
ROS Reactive oxygen species
RT Room temperature
s.c. Subcutaneous
SOC Super optimal broth with catabolite repression SSC Sideward scatter
TACE Tumor necrosis factor converting enzyme TAE Tris-acetate-EDTA
TNF-a Tumor necrosis factor a Treg Regulatory T cells
UPR Unfolded protein response
wt Wild type
µg microgram
1 Introduction
1.1. Purinergic receptors
Purinergic receptors are ligand-gated plasma membrane proteins expressed by almost all mammalian tissues that mediate several physiological and pathophysiological effects including neurotransmission, pain and inflammation (Burnstock 2007; Di Virgilio et al. 2018). These receptors are mainly activated by extracellular nucleotides, i.e., nicotinamide adenine dinucleotide (NAD) and adenine triphosphate (ATP). These nucleotides are released from the cells under physiological and pathological conditions (inflammation and cell death) (Kroemer et al. 2013).
Purine receptors belong to two subfamilies (Boeynaems et al. 2005), termed P1 and P2, that are activated by adenosine (nucleoside) or ADP/ATP (nucleotide), respectively. The P2 receptors are subdivided in P2X and P2Y receptors on the basis of structural differences and cellular functions, upon activation. P1 and P2Y are G protein coupled metabotropic receptors while the P2X subfamily belongs to ionotropic receptors (Burnstock 2007; Bilbao et al. 2012). The functional P2X receptors are able to form trimeric ion channels, either homomeric or heteromeric (Boeynaems et al. 2005; Burnstock 2007; Rissiek et al. 2015). Activation of P2X receptors results in influx of Ca2+ and Na+ ions and efflux of K+ ions. The P2X receptors differ in sensitivity to ATP from nanomolar (nm) to micromolar (µM) concentration (Burnstock 2007).
1.1.1. P2X7 ion channel
The P2X7 receptor is predominantly expressed by various immune and tumor cells. P2X7 is a unique member within the P2X family, structurally by the presence of a long intracellular carboxy-terminal tail that contains several signaling motifs, and biochemically by its requirement for comparatively high (upper micromolar to low millimolar) concentrations of ATP (North 2002). P2X7 channel consists of three subunits that self-assemble in homomeric trimer during translation and make stable complexes. The channel is composed of 595 amino acids which makes it the largest member of P2X family (Surprenant et al. 1996).
Gating of P2X7 usually occurs by extracellular ATP, but ADP-ribosyltransferase-C2 (ARTC2) expressing cells can also be gated by NAD-dependent ADP-ribosylation (Adriouch et al. 2002). The N-terminal domain of P2X7 is subject to alternative splicing. The mP2X7 (a) and (k) isoforms differ in the N-terminal 42 (a) and 39 (k) amino acids comprising the N terminal domain and most of the first transmembrane domain (Fig. 1A). In addition, C57BL/6 mice distinguish from Balb/c and wild-type mice by a functionally relevant polymorphism in the cytosolic domain (L451P) (Schwarz et al. 2012).
Figure 1. P2X7 structure, splice variants and mutants: P2X7 structure consists of
homotrimer sub units, with two transmembrane domains connected with an extracellular domain. The two mouse P2X7 (mP2X7) splice variants (a) and (k) differ at the first transmembrane domain with 42 amino acids in (a) isoform replaced with 39 amino acids in (k) isoform (Fig A & B). C57BL/6 mice differ from Balb/c mice with a single amino acid at 451 position where leucine in C57BL/6 mice is replaced with proline in Balb/c (Fig.1A) (Schwarz et al. 2012).
1.2. Purinergic signaling
ATP release from cells can serve as a paracrine signal for intercellular functions. ATP is present at millimolar concentrations within the cells, but only in nanomolar concentrations in the extracellular milieu. Under steady-state conditions, activation of
A
ATP concentration reaches >60 µM within 2 min after cell stimulation (Yip et al., 2009). These ATP concentrations are sufficient to stimulate P2X receptors (Klapperstück et al. 2001; North 2002; Gever et al. 2006). Following its release, ATP hydrolyzes gradually to ADP, AMP and adenosine by the action of ectonucleotidases (NTPDases and NPPs), ecto-5′-nucleotidase and alkaline phosphatases (Joseph et al. 2004; Robson et al. 2006; Zimmermann et al. 2012). This process prevents the desensitization of receptor and facilitates rapid and sustained purinergic signaling (Yegutkin 2008).
Figure 2. Purinergic signaling: Nucleotides such as adenosine triphosphate (ATP)
and nicotinamide adenine dinucleotide (NAD) are released from the cells in inflammatory conditions. ATP binds and activates the P2X receptors (mainly involved in proinflammatory responses) and NAD acts as a substrate for ARTC (ecto-ADP-ribosyl transferase) receptors. ATP is quickly degraded by Ectonucleoside triphosphate diphosphohydrolase-1 (NTPDase-1) to AMP which is further dephosphorylated to a potent anti-inflammatory and immune suppressor adenosine in the presence of ecto-5'-nucleotidase (5' NT). Modified from Stephan Menzel (unpublished figure).
Binding of the extracellular ATP to the P2X7 receptor induces a conformational change within milliseconds that allows the influx of extracellular Ca2+ and Na+ and the efflux of K+. Prolonged stimulation of P2X7 for seconds to minutes results in the formation of molecularly ill-defined pores that permit the passage of substrates up to 900 Daltons. This is accompanied by activation of membrane metalloproteases like ADAM17 (TACE, TNF-alpha converting enzyme) that cleave certain substrates such as the homing receptor L-selectin (CD62L) from the cell surface, as well as by the
phosphatidylserine (PS) exposure on the outer surface of the cell membrane as an early indicator of the induction of apoptosis (Haag et al. 2007).
Figure 3. P2X7 activation: P2X7 is an ATP-gated cation channel expressed in the
plasma membrane of many immune cells and tumors that have been linked to inflammation and tumor growth. Stimulation of P2X7 induces channel opening, influx of Ca2+ and Na+ ions; and efflux of K+ ions and the membrane pores formation that enable the influx of molecules up to 900Da. Upon P2X7 stimulation, activation of metalloproteases results in CD62L shedding from the cell surface as well as phosphatidylserine (proapoptotic marker) flipping from inner face of cell membrane to the outer membrane. Prolonged and strong stimulation of P2X7 leads to cell death (Scheuplein et al. 2009).
1.3. P2X7 in cancer
Beyond its importance for energy metabolism within the cell, ATP can be released from cells to play an important role as an extracellular signaling molecule under conditions of inflammation, tissue damage, and cell death (Pellegatti et al. 2008; Eltzschig et al. 2012). In tumor microenvironment, the concentration of ATP is remarkably high, which exerts multiple effects, including the conduction of survival signals as well as regulation of immune response via purinergic receptors (P2X or P2Y) (Seman et al. 2003; Di Virgilio et al. 2018). Among these, the P2X7 receptor has gained attention because it is expressed on a wide range of hematopoietic cells and also by many tumors (Di Virgilio et al. 2009).
The expression of the P2X7 receptor by tumor cells can have ambivalent effects for the tumor. Strong and prolonged stimulation of P2X7 has cytolytic effects on many
cells, presumably, due to its capability to induce membrane pores. Nonetheless, under steady-state conditions, the expression of P2X7 reportedly has a trophic effect on tumor cells by stimulating cell metabolism and increasing ATP production (Adinolfi
et al., 2012; Di Virgilio, 2012).
1.4. P2X7 blocking and potentiating nanobodies
Nanobodies are heavy chain only antibodies from camelids (Danquah et al. 2016). They consist of smallest antigen binding domains that can be used for therapeutic purposes. These single domain antibodies have various advantages over conventional antibodies i.e., nanobodies can interact with conformational epitopes that are not available to conventional antibodies. The monomeric nanobodies have superior tissue penetration and rapid effect compared to antibodies (Wesolowski et al. 2009; Van Bockstaele et al. 2009; Muyldermans 2013). The half-life of nanobodies can be increased by fusing them with albumin-specific nanobody (Tijink et al. 2008).
13A7 HLE (half-life extended) nanobody is a potent blocker of ATP mediated P2X7 activation, whereas 14D5 potentiates the effects of ATP on the P2X7 receptor. The blocking of P2X7 helps in attenuation of various inflammatory conditions i.e., glomerulonephritis (Danquah et al. 2016).
1.5. Immunogenic cell death (ICD)
Cell death by the apoptosis or necrosis pathways has been known for decades. Under steady state conditions, cell death mainly occurs via apoptosis, which is either tolerogenic or ignored by the immune system. In last two decade, the concept of immunogenic cell death (ICD) has evolved, which refers to a regulated cell death pathway that is able of activating an adaptive immune response against dead cell antigens derived from cancer cells (Kroemer et al. 2013; Bezu et al. 2015; Serrano-del Valle et al. 2019). Such antigens allow a strong antitumor immune response when presented to T cells (Chen and Mellman 2013).
A series of events is required for the effective removal of tumor cells by anticancer immune response. At the start, the released neoantigens are taken by dendritic cells (DCs) for processing. This may be accompanied by the release of danger signals such
as proinflammatory cytokines or the liberation of intracellular molecules. In the next step, captured antigens by DCs are presented on major histocompatibility class I (MHCI) and major histocompatibility class II (MHCII) molecules to T cells, which leads to priming and activation of effector T cell response against cancer specific antigens(Chen and Mellman 2013). At this stage, the type of immune response is determined by the ratio of effector T cell vs regulatory T cells (Motz and Coukos 2013). At the last stage, if the ratio of effector T cells is higher than the regulatory T cells, the cytotoxic T cells infiltrate the tumor bed and kill the target cancer cell (Zhu and Lang 2016).
Four different types of ICD have been described to date, each associated with a specific type of stimulus and danger signal. 1) pathogen driven ICD, which is one of the shielding mechanisms against pathogens; 2) necroptosis, an alternative form of regulated cell death, which requires RIPK3, a well-known regulator of inflammation in health and disease, 3) ICD-induced by physical signals such as irradiation, Hyp-PDT (Hypericin-Photodynamic-therapy) and HHP (High hydrostatic pressure) and 4) some chemotherapeutic drugs i.e., anthracyclines also induce immunogenic cell death by targeting ER stress and release of reactive oxygen species (Apetoh et al. 2007; Obeid et al. 2007; Michaud et al. 2011; Zitvogel et al. 2015).
Several mediators of immunogenic cell death have been well-known. These include the exposure of endoplasmic reticulum (ER)-resident protein calreticulin (CRT) on the cell surface, the cell death-associated release of high-motility group box 1 (HMGB1) protein, and the secretion of ATP (Panaretakis et al. 2009; Michaud et al. 2011; Garg et al. 2012). Not all chemotherapeutic drugs induce calreticulin translocation, ATP secretion and release of HMGB1. Doxorubicin (DOX) is an anthracycline produced by
Streptomyces peucetius. The structure of DOX contains a planar aromatic
chromophore portion that intercalates between two base pairs of the DNA. As a result, replication is stopped and DOX-treated cells die. Due to its fluorescent activity, the accumulation of this chemotherapeutic drug in cells can be detected using flow cytometry. DOX is used as a common treatment against cancer of the hematopoietic system as well as cancer of internal organs (Gewirtz 1999). DOX induces
immunogenic cell death by a mechanism that involves the release of ATP by dying tumor cells (Kroemer et al. 2013).
Figure 4:Immune activation by chemotherapeutic drugs (i.e., DOX): Increase in
endoplasmic reticulum (ER) stress by chemotherapeutic drugs may lead to ATP and high mobility group box 1 protein (HMGB1) release, and calreticulin (CRT) exposure on cell surface. CRT, ATP and HMGB1 bind to receptors on immature DCs, initiating the maturation and recruitment of DCs and antigen presentation to T cells. This process can lead to strong immune responses that can help in killing of chemotherapy resistant cancer cells. Modified from Kroemer et al., 2013.
Immunogenic cell death is regulated by two types of stress: ER stress and autophagy. Chemotherapy-induced ER stress results in the translocation of CRT to the cell surface from the ER lumen, and this occurs before the cell undergoes apoptotic cell death and marked by the exposure of phosphatidylserine on outer leaflet of cell membrane (Panaretakis et al. 2009). Calreticulin on outer leaflet of cell membrane acts as a strong phagocytic signal and enables dendritic cells to engulf stressed and dying tumor cells (Gardai et al. 2005; Chao et al. 2010). In immunocompetent mice, blockade of cell surface exposure of CRT negatively regulates the antitumor response of anthracyclines (Panaretakis et al. 2008). Therefore, ER stress-mediated immunogenic signal is necessary for the induction of anticancer immune responses by chemotherapeutic drugs (Obeid et al. 2007; Tesniere et al. 2010).
1.5.1. Unfolded Protein response
Tumor cells are continuously exposed to various stress factors i.e, environmental (hypoxia or nutrient deficiency) and genetic (genetic alteration, oncogene activation) factors. These factors are conventional initiators of ER stress (Urra et al. 2016). To deal with this stress, cells activate a well-defined process called the unfolded protein response (UPR). The UPR is controlled by three membrane-bound ER sensors, inositol-requiring enzyme-1 (IRE1), protein kinase R-like endoplasmic reticulum kinase (PERK), and activating transcription factor (ATF6) (Liu et al. 2015). Under steady state conditions, these sensors bind to BIP (binding immunoglobulin protein), which keeps them in an inactive form. Under stress conditions, BIP dissociates from the sensors and allow their self-activation and oligomerization. The activation of sensors help to restore the cell homeostasis by protein synthesis inhibition, degradation of misfolded proteins, and increase levels of redox enzymes that help in enhancing protein folding capacity (Szegezdi et al. 2006; Garg et al. 2012; Walter et al. 2018).
If ER stress sustains, the UPR can elicit a pro-apoptotic response controlled mainly by the IRE1 and PERK pathways. Activation of PERK results in phosphorylation of the eIF2a (eukaryotic initiation factor 2 a), which leads to inhibition of global protein translation to reduce protein load. However, some transcripts, like activating transcription factor 4 (ATF4), are translated proficiently in the presence of ER stress. Increased expression of ATF4 activates downstream pathways that involve the expression of genes responsible for amino acid and redox metabolism and apoptosis. ATF4 and the transcription factor CHOP (CAAT/enhancer-binding protein (C/EBP) homologous protein) are the vital factors of ER stress- mediated cell death (Serrano-del Valle et al. 2019). Activation of P2X7 receptor leads to phosphorylation of eIF2a,
induction of CHOP, and cleavage of caspase 3 with subsequent apoptosis (Chao et al. 2012).
Figure 5: Unfolded protein response: Under endoplasmic reticulum (ER) stress
conditions, the unfolded protein response (UPR) is activated which causes the activation of three branches of UPR: 1) PERK which results in phosphorylation of eIF2a, 2) IRE1 which activates the X-box binding protein (XBP1), and 3) activation of ATF6 to recover from this stress. The consequence of UPR response increases the protein folding, degradation of ER and transport proteins and shuts down the global mRNA translation. If the stress is not resolved, it leads to cell apoptosis. PERK, PRK like ER kinase. IRE1, Inositol-requiring protein 1. CHOP, C/EBP homologous protein; ROS, reactive oxygen species; XBP1s, transcriptionally active XBP1; XBPIu, unspliced XBP1. GAAD34, growth arrest and DNA damage-inducible protein 34 (Wang and Kaufman 2014).
1.5.2. Autophagy
Autophagy is also mandatory for immunogenic cell death. It does not interfere with the translocation of CRT and HMGB1 release, but it does positively affect the secretion of ATP. Only autophagy competent tumors recruit the DCs and T lymphocytes in response to chemotherapy. Inhibition of extracellular ATP degrading enzymes or high concentrations of extracellular ATP can help to restore a chemotherapeutic response in autophagy-deficient cancers. Therefore, autophagy plays a crucial role in the immunogenic cell death by releasing ATP (Michaud et al. 2011; Kroemer et al. 2013).
2 Aims of the project
It is still unclear whether activation of P2X7 under the conditions of chemotherapy will be beneficial to the tumor or the host. The aim of my project is to study the effects of P2X7 gating in different cancer cell lines (Yac-1, 4T1 and A20 cells) under treatment with the chemotherapeutic agents that induce immunogenic cell death (doxorubicin and Bortezomib).
This includes the following specific objectives:
• To assess the impact of P2X7 expression and gating for the growth and survival of tumor cells
• To evaluate the impact of P2X7 gating on tumor cells in the presence of chemotherapy
• To evaluate the consequences of P2X7 expression in tumor cells in-vivo by generating mouse models
3 Results
The results are divided into two sections. In section 3.1, the in-vitro synergistic cytotoxic effects of DOX in the presence of P2X7 receptor activation are presented. Additionally, the possible mechanisms i.e., enhanced ER stress and drug uptake are also shown. In section 3.2 the in-vivo role of P2X7 on tumor cells is presented, showing that expression of P2X7 (k) by tumor cells slowed down the tumor growth compared to tumors expressing non-functional P2X7 or treated with P2X7 blocking nanobody (13A7 half-life extended). Moreover, treatment of solid tumors with P2X7 potentiating nanobody (14D5 half-life extended) further reduced the tumor growth.
3.1. In-vitro results
3.1.1. P2X7 expression in Yac-1 cells
P2X7 receptors are expressed by most of immune and tumor cells. Here I investigated the expression of P2X7 on the murine lymphoma cell line Yac-1 by flow cytometry using anti-mouse P2X7 specific antibody. The anti-human P2X7-specific antibody was used as an isotype control. Results showed that Yac-1 cells have noticeable surface expression of P2X7 (Fig. 6A). To find out the splice variant of P2X7 in Yac-1 cells, DNA from Yac-1 cells was amplified by using P2X7 (a) and (k) specific primers (Fig. 6B).
Yac-1 cells express the (k) variant, which is mainly expressed by T cells and is more sensitive to ATP compared to the (a) variant (Schwarz et al. 2012).
Figure 6. Cell surface expression and splice variants of P2X7 in T cell lymphoma
(Yac-1): A) To check the expression of P2X7 on Yac-1 cells, cells were stained with fluorochrome-conjugated anti-P2X7 monoclonal antibody RH23-A44-Alexa 647 (shaded histogram) or anti human antibody L4, an isotype control (open histogram) and analyzed by flow cytometry. B) cDNA from Yac-1-cells was amplified by PCR
using primers specific for the (a) or (k) splice variants of P2X7 (lanes 3 and 4). Plasmids encoding P2X7 (a) or (k) were used as positive controls (lanes 1 and 2).
3.1.2. P2X7-mediated CD62L shedding, PS flashing and DAPI uptake
It has previously been shown that short-term (20-30 min) treatment of P2X7-transfected HEK cells with ATP resulted in P2X7 mediated CD62L shedding, translocation of phosphatidylserine (PS) to the outer surface of the cell membrane, and low-level uptake of DAPI as an indicator of pore-formation (Schwarz et al. 2012). To determine whether Yac-1 cells also exhibit similar responses upon ATP treatment, cells were incubated with various concentrations of ATP (50 µM - 800 µM) for 30 min, and samples were measured by flow cytometry. CD62L shedding (Fig. 7A-B) and PS
translocation (Fig. 7C) were observed at relatively low concentrations of ATP (50µM),
while DAPI uptake required ATP concentrations > 200 µM (Fig. 7D).
Figure 7. P2X7R mediates ATP-induced CD62L shedding, PS exposure, and dye
uptake: A-D) Yac-1 cells were stained with an anti-CD62L antibody and DNA-staining
dye, and incubated for 30 min at 37°C in the presence of the indicated concentrations of ATP. Cells were washed with annexin-V buffer and stained with annexin-V antibody to detect the surface exposure of phosphatidylserine (PS) for 15 min at room temperature in dark. Samples were analyzed by flow cytometry and mean fluorescence intensity (mfi) was calculated by flowjo software. ATP induced CD62L shedding (A-B), PS exposure (C), and DAPI uptake (D). Data presented as mean ± SEM
3.1.3. ATP-mediated cytotoxicity is blocked by P2X7 antagonist 13A7
It has been demonstrated that the prolonged activation of P2X7 by ATP results in cell death (detected by propidium iodide staining) which was prevented by the P2X7 blocking nanobody 13A7 (Farrell et al. 2010; Danquah et al. 2016). I investigated here whether this effect can also be observed in 1 cells. Overnight incubation of Yac-1 cells at 37 °C with the indicated concentrations of ATP resulted in cytotoxicity with an EC50 around 300 μM. ATP cytotoxicity was mediated by the P2X7 receptor, since pre-incubating the cells with the P2X7-specific nanobody 13A7 blocked this effect. The inhibitory effect is dose dependent as it is overcome at high concentrations of ATP Fig. 8A-B). Cells were also stained for annexin V to check the apoptotic cell
death via phosphatidylserine (PS) flashing. Interestingly, P2X7 activation did not only induce necrotic cell death that is shown as DAPI and annexin V double positive cells but also induce apoptotic cell death, annexin V only positive cells (Fig. 8A). The
presence of P2X7 blocking nanobody more strongly blocks apoptotic effects (Fig. 8C)
compared to overall cell death (Fig. 8B).
Figure 8. ATP induced toxicity in Yac-1 cells: A-C) Yac-1 cells were preincubated
with the P2X7 blocking nanobody 13A7 for 15min at room temperature on a roller in 500 µl of RPMI complete medium, and incubated for 24 h at 37 °C and 5% CO2 with the indicated concentrations of ATP. After washing with annexin V buffer cells were stained with annexin V for 15 min at room temperature in the dark. To stain dead cells, DAPI was added immediately before flow cytometry analysis. A) Representative FACS
plots for DAPI uptake vs annexin V staining in untreated samples (ø) and 300µM ATP treated samples in the presence and absence of P2X7 blocking nanobody 13A7. B)
XY graph shows the dose response curve for ATP induced cell death which is blocked by 13A7 in a dose dependent manner. C) Mean fluorescence intensity of annexin V in
ATP ± 13A7 treated samples is shown. The experiment was performed in triplicate for each condition. This figure is representative of three independent experiments with similar results. Statistics: two-way analysis of variance; posttest: Bonferroni test. ***, p<0.001.
3.1.4. Gating of P2X7 enhances the uptake and cytotoxicity of doxorubicin in
Yac-1 lymphoma cells
Doxorubicin (DOX) is an anthracycline drug used in the treatment of lymphomas and various solid tumors (Gewirtz 1999). To assess the effect of P2X7 gating on DOX cytotoxicity, Yac-1 cells were incubated overnight with various concentrations of DOX in the presence of 200µM ATP (Fig. 9A-B). While DOX at a concentration of 30 nM
induced death in about 8% of cells, the combination with 200 µM ATP increased cell death to above 80%. This increase in cell death was completely blocked by pre-treating the cells with the nanobody 13A7. Importantly, gating of P2X7 by 200 µM ATP in the absence of DOX caused only low levels of cell death (15-20%).
Figure 9. Gating of P2X7 enhances the cytotoxicity of doxorubicin in Yac-1 cells:
The cytotoxic effects of ATP and DOX on Yac-1 cells after overnight incubation were determined by flow cytometry. A-C) Cells were incubated with 200 µM ATP and
varying concentrations of DOX in the presence or absence of the P2X7-inhibitory nanobody 13A7. To assess vitality, DAPI was added immediately before analysis. Cell death was defined as the uptake of DAPI combined with a reduction in forward scatter as shown in A). Representative FACS plots and bar graph are shown in Fig. A and B
Error bars represent standard deviations from the mean. The results are representative of at least three independent experiments.
3.1.5. Doxorubicin cytotoxicity in Yac-1 cells is independent of P2X7 activation.
Fig. 9 showed that Yac-1 cells are sensitive to DOX-induced toxicity. Twenty-four
hours treatment is enough to kill the cells with EC50 150nM (Fig. 9B) and apoptosis
is the dominant pathway of DOX-induced cell death (Fig.10 A and C). To investigate
the role of P2X7 in DOX-induced toxicity, cells were preincubated with P2X7 blocking nanobody 13A7. Results showed that blocking of P2X7 did not affect the DOX cytotoxicity (Fig. 10A-C).
Figure 10. Doxorubicin cytotoxicity: Yac-1 cells were preincubated with 13A7 for
15 min on a roller. Cells were treated with the indicated concentrations of Doxorubicin ± 13A7 overnight. Cells were then washed with annexin V buffer and stained for the exposition of PS with APC conjugated annexin V for 15 min in the dark at room temperature. The DNA-staining dye DAPI was added 2 min before analysis. Data collection was carried out by flow cytometry. A) Representative FACS plots for DAPI+ vs annexin V+ cells. B) Dose response curve for DOX toxicity in the presence and absence of 13A7. C) Summary of percentage of annexin V+ cells.
3.1.6. Gating of P2X7 enhances uptake of doxorubicin at early timepoints.
Activation of the P2X7 receptor is known to induce the pore formation that allows passage of substances up to a molecular weight of 900 Dalton through the plasma membrane (Surprenant et al. 1996). I therefore considered the possibility that gating of P2X7 might enhance the cytotoxicity of DOX by facilitating the entry of DOX into the cells. Since DOX is a fluorescent compound, its content in cells can be determined conveniently by flow cytometry. Thus, the effect of ATP on DOX uptake was examined. Indeed, ATP at concentrations as low as 100 µM facilitated the entry of
DOX into the cells within 1 h (Fig. 11A). However, at the end of the overnight assay
the total DOX content was only marginally increased in the presence of ATP (Fig. 11B).
Figure 11. Gating of P2X7 enhances uptake of doxorubicin at early timepoints: A)
Yac-1 cells were incubated for 1 and 4 h with 1 µM DOX, ATP (100 and 200 µM), or a combination of both. The accumulation of fluorescent DOX was monitored by flow cytometry. B) Yac-1 cells were treated overnight with increasing concentrations of
DOX in the presence or absence of 200 µM ATP and the P2X7 inhibitory nanobody 13A7. The DOX uptake was measured by FACS. Statistics: Two-way analysis of variance; posttest: Bonferroni test. ***, p<0.001.
3.1.7. ATP-induced cell death is rapid, but death due to ATP/DOX synergism is a late event
Signaling through P2X7 induces rapid cell death in primary T lymphocytes that is apparent within 30 min (Seman et al. 2003), presumably associated with the opening of the large non-selective membrane pore (Di Virgilio et al. 2001). By contrast, apoptosis mediated by DOX is a slow process requiring intracellular signal transduction and de novo synthesis of pro-apoptotic proteins (Gewirtz 1999; Peidis
et al. 2011). I asked whether cell death caused by the synergistic interaction of ATP
and DOX occurred early or late after exposure to the two substances. Using low concentrations of ATP just above the threshold for ATP-mediated cell death (100 and 200 µM), and a high concentration of DOX (1 µM) that is highly cytotoxic (see Fig. 10),
I investigated the effects of ATP and DOX 60 min after exposure to the two substances (Fig. 12A). The results confirmed that at one-hour low-dose ATP had killed some of
the cells, while even high-dose DOX had no cytotoxic effect. I then followed the time course of cytotoxicity, using concentrations of ATP (100 µM) and DOX (30nM) that were at or just below the threshold for inducing cell death by themselves (Fig. 12B).
detectable within the first four hours after incubation. At 16 hours, however, this effect was no longer visible, presumably due to degradation of ATP and proliferation of cells in this time period. No cytotoxicity was observed at any time-point for 30 nM DOX. Of note, the presence of 30 nM DOX did not add to the toxicity of ATP during the first four hours. However, after 16 hours of incubation, the combination of ATP and DOX killed 80% of the cells. Together, these observations suggest that gating of P2X7 strongly enhances DOX-mediated death, but that the presence of DOX does not affect the early cytotoxicity mediated by ATP.
Figure 12. ATP-induced cell death is rapid, but death due to ATP/DOX synergism is a late event: A) Yac-1 cells were treated with 1 µM DOX in the presence or absence
of 100 or 200 µM ATP for 1 h. Cell death was determined by adding DAPI before analysis by flow cytometry. B) Cells were treated with 100 µM ATP, 30nM DOX, or a
combination of both for 2, 4 and 16 h. Cell viability was determined by staining with DAPI. Data is shown as mean± SEM. Statistics: two-way analysis of variance; posttest: Bonferroni test. *, p<0.05, ***, p<0.001.
3.1.8. Transient activation of P2X7 marks cells for doxorubicin-induced death
The signal transmitted by gating of P2X7 is rapid, and the ligand ATP is not stable for long periods of time in the extracellular milieu. By contrast, DOX accumulates within cells, and presumably DOX signaling for death is the result of a long process. I wondered whether the enhancement of DOX-mediated cell death by ATP required long-lasting signals at the P2X7 receptor. To address this question, Yac-1 cells were co-incubated with ATP and DOX for varying periods of time between 1 h and 16 h, then washed and resuspended in fresh medium for further culture overnight. I found that a 1h co-incubation with ATP and DOX sufficed for cytotoxic synergism, i.e. to induce death of > 80% of cells after further incubation overnight, as evidenced by reduction in FSC signals and intense staining with DAPI. In contrast, exposure of cells
to ATP or DOX alone for one hour caused only low levels of cell death (Fig. 13A and C).
At the same time, I monitored the uptake of DOX into the cells (Fig. 13B). The results
show that the presence of ATP significantly increased the incorporation of DOX into the cells
Figure 13. Transient gating of P2X7 marks cells for doxorubicin-induced death:
A-C) Yac-1 cells were incubated for 1 h with 500 nM DOX, 100 µM ATP, 200 µM ATP,
or the indicated combinations of ATP and DOX. The cells were then washed and cultured overnight in complete RPMI medium. Cell death and DOX accumulation were determined by flow cytometry. A-C) Representative FACS plots and bar graph
showing cell death in relationship to the duration of treatment with DOX and/or ATP.
B) DOX uptake by cells treated for 1h. This experiment is representative of three
separate experiments with similar outcome. Data shown as mean± SEM. Statistics: two-way analysis of variance; posttest: Bonferroni test. ***, p<0.001
3.1.9. Synergistic effects of P2X7 signaling on doxorubicin toxicity are
dependent on the P2X7 isoform
In the first phase of the project, I found that Yac-1 cells were sensitive to ATP and DOX with EC50s of 300 µM and 200 nM, respectively. Treatment with ATP increased the cytotoxic effects of DOX in a synergistic manner. To further support our results, we cloned P2X7 splice variants (a) and (k) into the retroviral vector pMXs. Retroviral vectors have the ability to transform their single strand RNA to double stranded DNA that stably incorporates into the target cell genome (Anson 2004).
4T1 cells are a mouse mammary carcinoma cell line. Various characteristics of 4T1 cell line make it suitable for in-vivo studies of breast cancer. It is easily transplantable, highly tumorigenic and invasive. Metastatic spread of 4T1 cells is very similar to that of human mammary cancer (Pulaski et al. 2000a). In recent studies, 4T1 cell line was also used to evaluate immunotherapy strategies focussed on tumor specific CD8 and CD4 lymphocytes (Pulaski and Ostrand-Rosenberg 1998; Pulaski et al. 2000b). Therefore, I used 4T1 cell line to check the role of P2X7 in tumor cells.
To see if P2X7 also synergistically enhances the effects of DOX in this cell line, I transduced the breast tumor cell line with the P2X7 (a) and (k) splice variants, as well as with a non-functional variant of P2X7, which has a point mutation at position 294 that makes it insensitive to ATP (Adriouch et al. 2009). FACS analyses revealed a high level of P2X7 expression in 4T1 cells transfected with both P2X7 variants and the dead mutant when stained with the monoclonal antibody RH23-A44-Alexa 647, which is specific for murine P2X7 (Fig.14A). I treated the 4T1 wild type and P2X7-transduced cell lines with increasing concentrations of DOX in the presence of 300 µM ATP for 48 h. Interestingly, I found that ATP increased the cytotoxic effects of DOX in the 4T1 P2X7 (k) cell line in a similar fashion as in Yac-1 cells (Fig. 14D), while the 4T1 P2X7 (a) cells did not show any synergism between DOX and ATP at this concentration (Fig. 14C).
Figure 14. The synergistic effect of P2X7 on doxorubicin cytotoxicity is
dependent on the P2X7 isoform: A) 4T1 mammary carcinoma cells were retrovirally
transduced with P2X7 (a) and (k) splice variants, as well as with a non-functional P2X7 variant (dead mutant). B-E) cells were treated for 48 hours with various concentrations
of DOX and 300 µM ATP in the presence or absence of 13A7. Cell death was measured by the acid phosphatase assay. Histograms in Fig. A show the expression
of P2X7 in different 4T1 cell lines. 4T1 wt (gray) does not express any P2X7, while 4T1 cells transduced with P2X7 (a) (purple), P2X7 (k) (blue) and P2X7 R294A (turquoise) show the similar expression levels of P2X7. Dose response curves for DOX cytotoxicity in the presence or absence of ATP are shown in 4T1_wt (B), 4T1 P2X7 (a) (C), P2X7
3.1.10. P2X7 (k)-expressing A20 B lymphoma cells also show enhanced
doxorubicin cytotoxicity in the presence of ATP
I generated a further model system by transducing the A20 B lymphoma cell line with the P2X7 (a) and (k) splice variants. The rationale for doing so was twofold. Firstly, A20/P2X7 provides a second lymphocyte model that allows for direct comparison of P2X7 variants; and secondly, as a B cell-derived tumor they might provide a good model to study the unfolded protein response. FACS analyses showed a high level of P2X7 expression in A20 cells transfected with both P2X7 variants (a) and (k) when stained with RH23-A44-Alexa 647 (Fig. 15A). Therefore, this cell line is appropriate to
investigate P2X7 signaling in tumor cells as well as to find out the role of the two splice variants in the DOX/ATP synergism. I tested all the cell lines for ATP sensitivity. 30 min treatment with different concentrations of ATP showed that A20 P2X7 (k) is very sensitive to ATP (IC50: 300µM) and PS exposure on living cells (gated on DAPI negative cells) was observed at 100 µM ATP (Fig. 15B-D). A20 P2X7 (a) cells were much less sensitive to ATP (IC50: 1mM). A20 wt did not respond to ATP treatment. I also tested all the cell lines for DOX sensitivity in presence of ATP and the P2X7 blocking nanobody 13A7. I observed enhanced DOX sensitivity only in P2X7 (k) transduced cells in the presence of 200 µM ATP (Fig. 15E-G). A20 wt and A20 cells
transfected with P2X7 (a) variant did not show any additive toxic effects. I have also used higher concentration of ATP in the presence of DOX to check if A20 P2X7 (a) shows any synergistic effects, and I observed that P2X7 (a) required more than 500µM ATP for synergistic toxic effects (data not shown).
Figure 15. P2X7 splice variants differ in their effects on synergistic cytotoxicity:
A) The B lymphoma cell line A20 was transduced with the two P2X7 splice variants (a) and (k) and the expression was checked by flow cytometry as explained in figure 14. A20-wt does not express P2X7 (gray shaded histogram), P2X7 (a) and (k) transduced A20 cells showed similar expression of P2X7 as shown in orange shaded (P2X7 (a)) and blue shaded (P2X7 (k)) histograms. B-C) 30min incubation with various
concentrations of ATP showed that P2X7 (k) has a higher sensitivity to ATP than P2X7 (a), while un-transfected cells were completely insensitive to ATP. B) Representative
FACS plots for cell death after 30 min and summary of Fig. B is shown C. D) PS
flashing (increase in mean fluorescence intensity (mfi) of annexin V) in un-transfected and P2X7 transfected A20 cells. E-G) Acid phosphatase assay was performed to
check the cell viability in all A20 cell line treated with DOX +/- ATP and 13A7. The presence of ATP (300µM) did not affect DOX induced cell death in A20_wt (E) and
A20 P2X7 (a) cell line (F). LC50 for doxorubicin was significantly reduced in A20/P2X7 (k) cells treated with 200 µM ATP (G).
3.1.11. P2X7 and doxorubicin co-operate in the induction of
phagocytosis-promoting signals
Anthracycline drugs are known inducers of type 1 immunogenic cell death (Garg et al. 2016). A hallmark of ICD is the translocation of calreticulin (CRT) from the ER to the cell surface. Cell surface CRT provides a phagocytosis signal to APCs, which together with the release of danger signals stimulates the immune reaction against the dying cell (Kroemer et al. 2013). On the other hand, P2X7 signaling causes the translocation of phosphatidylserine (PS) to the exterior leaflet of the plasma membrane, which facilitates the recognition and uptake of the cell by professional phagocytes (Adriouch et al. 2002). I therefore asked whether P2X7 and DOX also co-operate in the induction of these phagocytosis-promoting signals. A six-hour culture in the presence of 100 µM ATP did not induce the exposure of PS or CRT on live cells. Co-incubation with the P2X7-enhancing nanobody 14D5 (Danquah et al. 2016) led to a small increase in PS, but not CRT exposure. DOX alone (300 nM) induced exposure of both PS and CRT, which was not affected by co-incubation with 100 µM ATP alone, but strongly enhanced in the presence of ATP + 14D5 (Fig.16 B). Incubation with DOX
also caused some cell death, which was also slightly increased in the presence of ATP and 14D5 (Fig. 16A). Interestingly, incubation with DOX alone for 6 hours caused
an increase in membrane permeability in living cells, as shown by an increase in low-level DAPI uptake (Fig.16A).
Figure 16: P2X7 and doxorubicin co-operate in the induction of
phagocytosis-promoting signals: Yac-1 cells were incubated for 6 h with 300 nM DOX and various
concentrations of ATP (0 - 100 µM) in the presence or absence of the P2X7-enhancing nanobody 14D5, and analyzed for cell vitality (A), binding of annexin V or surface
expression of calreticulin (B) by flow cytometry on a FACS Canto2. In B live cells are
shown. Staining for calreticulin (CRT) was performed with unlabeled rabbit monoclonal antibody (clone D3E6) followed by secondary staining with AF647-coupled anti-Rabbit-Ig antiserum. Annexin V staining was done by incubating the cells with annexin V antibody for 30 min at room temperature in dark. Cell vitality was determined by co-staining with DAPI (A). Mean fluorescence intensity of annexin V
and calreticulin are shown in C and D respectively. Statistics: two-way analysis of
I further confirmed the increase in ER stress in the presence of ATP and DOX by microscopy. Fibroblast cell line 3T3 transfected with P2X7 (k) was further stably transduced with calreticulin fused to GFP. The expression of CRT was checked by flow cytometry and under the microscope. Cells were treated with DOX, ATP or combination of both for 4 h. A decrease in CRT at the ER was observed in the combined therapy (Fig. 17 D) as well as stress granules were formed in the treated
cells compared to control.
Figure 17. Enhanced ER stress in fibroblast cell line: A-D) To check the calreticulin
surface translocation, P2X7 (k) transfected 3T3 cells that were retrovirally transduced with calreticulin eGFP were seeded (5*104/ sample) on cover slips (previously treated with 0.1 mg /ml poly L lysin (PLL), treated with 500 nM DOX in the presence and absence of 200 µM ATP for 4 h. Cells were permeabilized with 1% Triton for 5 min and blocked with blocking buffer (1% BSA and 0.1 % triton) for 40 min at room temperature and stained with Phalloidin (1:50) for 30min to visualize the actin cytoskeleton. Cover slips were mounted with mounting medium containing DAPI on slides. Stress granule formation was observed by using confocal Microscope (sp5). Untreated sample (ø) showed intact cytoskeleton (red) and nucleus (blue). No surface translocation of calreticulin (green) was observed (A). Treatment with ATP alone (B)
and in the combination with DOX (D) resulted in stress granule formation. Fig. C
3.1.12. P2X7 also enhances the cytotoxicity of Bortezomib, a known
inducer of the unfolded protein response
Surface exposure of calreticulin induced by DOX or other mediators of immunogenic cell death results from ER stress and consequent triggering of the unfolded protein response (UPR) (Garg et al. 2012). I therefore investigated the effect of P2X7 gating on the UPR induced in A20 cells by overnight treatment with Bortezomib (BTZ). BTZ is a known type 1 immunogenic cell death inducer that causes protease inhibition in the ER, resulting in ER stress and triggering of UPR (Ri 2016). I evaluated the survival of A20 cells transduced with P2X7 (k) in the presence or absence of ATP (Fig. 18 A-B).
Treatment with 3 nM BTZ alone caused death in approximately 10 % of cells, while only few cells died in response to 100 µM ATP alone. However, in the presence of 100 µM ATP, 3 nM BTZ was toxic for more than 60% of the cells. This effect was blocked entirely by pre-treatment with the P2X7-antagonistic nanobody 13A7.
Figure 18. P2X7 activation results in enhanced Bortezomib toxicity: A-B) A20/P2X7
(k) cells were treated overnight with different concentrations of Bortezomib (3-100 nM) or in the combination of 100 µM ATP in the presence or absence of 13A7 nanobody. DAPI was added immediately before analysis to stain the dead cells. Cell viability was assessed by flow cytometry. Representative FACS plots for cell death and summary of data are shown in A and B respectively.
3.1.13. P2X7 and doxorubicin synergize to activate Caspase-3
It has been reported that activation of caspase-3 is an important step in the pathway that mediates apoptotic cell death resulting from ER stress (Hitomi et al. 2004). I therefore investigated the role of P2X7 signaling in caspase-3 activation. Treatment of Yac-1 cells with either 200 µM ATP or 300 nM DOX for 6 h induced activation of caspase-3 in live cells (Fig. 19A-B). This was supra-additively enhanced when the
two compounds were given together. The synergistic effect of ATP was blocked by pre-incubation of the cells with the P2X7-inhibitory nanobody 13A7. Staurosporin (STR), a known inducer of caspase-3-dependent apoptosis was used as a positive control (Fig. 19C).
Figure 19. P2X7 and doxorubicin synergize to activate caspase-3: A-B) Yac-1 cells
were incubated at 37°C for 6 h with 200 µM ATP, 300 nM DOX, or a combination of both, or 300 nM staurosporin (STR), washed, and stained with the FLICA caspase-3 kit to detect activated caspase-3. DAPI was added immediately before analysis to monitor cell death. A) Representative FACS plots for cell viability. B) Mean
fluorescence intensity (mfi) for caspase 3 in live cells (DAPI negative cells). C) STR
was used as a positive control for caspase-3-dependent apoptosis.
3.1.14. Gating of P2X7 alone results in Calreticulin (CRT) translocation
The previous experiments established that P2X7 synergistically co-operates with DOX in the induction of the ER stress/UPR pathway. In light of these observations I asked whether gating of P2X7 by itself was also sufficient to trigger the UPR. To do so, I investigated whether ATP alone could induce CRT translocation to the cell membrane
(k) cells and more than 1 mM ATP in the case of A20 P2X7(a) was sufficient to induce CRT surface exposure in a concentration-dependent manner.
Figure 20. P2X7 activation results in CRT surface translocation: A-D) A20/P2X7 (a)
or (k) cells were incubated with various concentrations of ATP for 30min. Cells were stained for calreticulin (CRT) using the unlabeled rabbit monoclonal antibody (clone D3E6) for 1 h on ice, followed by secondary staining with AF647-coupled anti-Rabbit-Ig antiserum for 30 min at 4 °C. Cell vitality was determined by co-staining with DAPI. Representative histogram for calreticulin translocation in A20 P2X7 (k) is shown in A
and summarized data in B. Calreticulin surface translocation in A20 P2X7 (a) cell line is shown in C and D.
3.1.15. P2X7-mediated CRT translocation, PS exposure, and pore
formation are regulated by the PERK pathway
The PKR-like ER kinase (PERK) is an important mediator of the immunogenic cell death pathway. Under conditions of ER stress, activation of PERK leads to phosphorylation of its substrate eukaryotic initiation factor-2a (eIF2a) and externalization of CRT (Panaretakis et al. 2009; Kepp et al. 2015). I therefore tested if P2X7-induced ER stress is regulated by PERK activation.
For this, cells were pre-treated with the PERK inhibitor GSK2606414 for 1 h and then incubated with ATP for short time periods (2-10 min). Interestingly, I found that, p-eIF2a, CRT translocation, PS exposure, and DAPI uptake induced by short-term
exposure to ATP were all blocked by inhibition of PERK (Fig. 21A-E). However, CD62L
shedding did not block by PERK inhibitor in the presence of 400 µM ATP (Fig. 21F).
Figure 21. P2X7-mediated DAPI uptake, PS flashing and CRT translocation are
blocked by inhibition of PERK: A-F) Yac-1 cells were pre-incubated with 10 µM of
the PERK inhibitor GSK2606414 (PERKi) for 1 h, and then cells were treated with 100 µM ATP for 10 min (A), 500µM ATP for 10 min (B & C) and 400 µM ATP for 2 min (D-F). Cells were washed and stained for p-eIF2a, surface CRT, PS exposure, CD62L shedding and DAPI uptake (explained in previous sections). Representative bar graphs show increase in phosphorylation of eIF2a (A), surface translocation of CRT
(B) and PS flashing (C) in ATP treated samples, all these affects are blocked in the
presence of PERK inhibitor. A representative FACS histogram for DAPI uptake is shown in Fig. D and summarized data in Fig E. Presence of PERK inhibitor blocks
ATP mediated DAPI Uptake. F) ATP mediated CD62L shedding did not affected in the
presence of PERK inhibitor. Statistics: two-way analysis of variance; posttest: Bonferroni. ***, p<0.001
3.1.16. Inhibition of PERK blocks the synergistic effect of P2X7 on
doxorubicin-mediated cytotoxicity
In previous experiments I found that PERK inhibition negatively regulates the P2X7 mediated effects i.e., DAPI uptake, CRT surface translocation and PS translocation. Here I investigated if the PERK pathway also affects the cytotoxic effects of DOX in the presence or absence of subthreshold eATP. Interestingly, inhibition of PERK had the opposite effects on the DOX-mediated cytotoxicity and the P2X7-mediated cytotoxic synergism (Fig. 22 A-B), showing that gating of P2X7 did not merely amplify DOX signaling by increasing its uptake.
Figure 22. Inhibition of PERK blocked the synergistic effect of P2X7 on
doxorubicin-mediated cytotoxicity: A-B) Yac-1 cells were treated with 10µM PERK
inhibitor (GSK2606414) for 1 h and then indicated concentrations of ATP, DOX or a combination of DOX and ATP were added overnight at 37°C and 5% CO2. Cells were washed and stained with DAPI immediately before analysis by flow cytometry. Each condition was performed in triplicate. A) Representation of cell viability (DAPI+ cells)
in the presence of different treatments with and without PERK inhibitor. B): Bar graph
representing the mean values ± SEM. Statistics: two-way analysis of variance; posttest: Bonferroni. ***, p<0.001
3.2. In-vivo results
3.2.1. Expression of P2X7 inhibits tumor growth.
Bioluminescence imaging (BLI) technology has become a powerful tool for real time monitoring of diverse molecules and tumor growth in-vivo, based on the generation of light by the luciferase/luciferin reaction. For in-vivo imaging of tumor growth, I transduced the Yac-1 cells with the Lentivirus LeGo-iG2-Puro+luc2+GFP, provided by Dr. Kristoffer Riecken, Center for Oncology, Interdisciplinary Clinic and Polyclinic for Stem Cell Transplantation, University Medical Center, Hamburg-Eppendorf, Germany), which encodes luciferase in combination with the enhanced green fluorescent protein (eGFP). To study the effect of P2X7 on tumor growth, luciferase transduced Yac-1 cells (that endogenously express P2X7) were injected intravenously into SCID mice, which were then treated with the P2X7-antagonistic nanobody 13A7-HLE (half-life extended) twice a week (Fig. 23A). Bioluminescence (BLI) imaging was
performed twice a week with the help of Mr. Michael Horn-Glander, IVIS Facility, University Medical Center Hamburg-Eppendorf, using small animal imaging system (IVIS-200, Caliper Life Sciences, Hopkinton, Massachusetts, USA). D-luciferin was injected in anesthetized mice 15 min before image acquisition. Mice (max. 5) were placed in imagining system and imaged. The total flux is calculated by photons per sec (p/s) in the region of interest (ROI) (Figure 23 B-C), by using Living image 4.2
software. Results showed that blocking of P2X7 resulted in accelerated tumor growth compared to placebo (Fig. 23 B-F).
Figure 23. Inhibition of P2X7 in Yac-1 cells promotes tumor growth: A) 2.5*105
Yac-1 cells were injected i.v. in SCID mice. At day 3, mice were randomized in 2 groups after in-vivo imaging using small animal imaging system (IVIS-200, Caliper Life Sciences). Mice were injected (i.p.) with PBS (placebo) or P2X7 antagonist nanobody, 13A7 HLE (50µg), twice a week. B-C) in vivo imaging of placebo and 13A7 HLE
injected mice at day 14. Images were captured 15 min after intraperitoneal D-luciferin injection. Red rectangles around mice in figure B and C represent the region of interest
(ROI). D) Comparison of tumor growth in 13A7.HLE- and placebo-injected mice is
shown as total flux (p/s). Data presented as mean ± SEM. E-F) Tumor size, measured as total flux (p/s),in individual mice treated with placebo (PBS) or 13A7 at day 17 and day 21.
3.2.2. Expression of functional P2X7R slows 4T1 tumor growth in-vivo
To confirm the role of P2X7 in tumor growth, I performed a similar experiment using the 4T1 breast tumor model with the assistance of Dr. Isabel Ben Batalla, Department of Oncology, University Medical Clinic, Hamburg-Eppendorf, Germany. As previously described, 4T1 cells were transfected with P2X7(k) or the non-functional mutant of P2X7 that is unable to bind ATP as a control. The 4T1 cells were injected into syngeneic Balb/c mice. Interestingly, 4T1/P2X7 (k) tumors grew significantly slower compared to controls (Fig. 24A), supporting the findings from in-vivo experiment
(Figure 23) showing that expression of P2X7 slows down tumor growth. The tumor samples were analyzed for T cell infiltration and apoptosis by histology. P2X7-expressing tumors exhibited higher degrees of T cell infiltration (CD3 staining) and apoptosis than the controls (Fig. 24 B-D)
Figure 24. Tumors carrying a functional P2X7R exhibit slower growth in-vivo: 4T1
cells (5*105) carrying either a non-functional P2X7 mutant (4T1) or P2X7(k) (4T1_P2X7) were injected into mammary fat pads of Balb/c mice. A) Tumor volume was assessed
in-vivo by caliper at the indicated time points. B-D) Tumor samples were stained for
The 4T1-injected mice were then treated with chemotherapy (DOX) when a tumor size of 100mm3 was reached. Significant tumor reduction was observed in DOX-treated control mice (Fig. 25 A). Surprisingly, DOX had no effect on the growth of
P2X7-expressing tumors (Fig. 25 B). Histological results showed that the presence of DOX
did not affect the infiltration of T cells in control and P2X7-expressing tumors (Fig. 25 C) while it significantly increased apoptosis in control group (Fig. 25 D).
Figure 25. Doxorubicin does not affect the tumor growth in P2X7 (k) expressing
tumors: A-B) 4T1 (P2X/ dead mutant) and 4T1(P2X7_k) cells (5*105) were injected in mammary fat pads of Balb/c mice and treated with 3mg/kg DOX at day 10 and 13. A-B) Tumor volume was assessed in-vivo assessed by caliper at the indicated time
points. C-D) For histology, tumor samples were stained for CD3 and caspase 3.
Statistics: two-way analysis of variance; posttest: Bonferroni. *, p<0.05, **, p<0.01, ***, p<0.001.
3.2.3. P2X7 enhancing nanobody slows down the tumor growth.
The tumor microenvironment contains a high concentration of extracellular ATP (eATP). It has been shown that eATP levels in tumors are in the range of 100-500µM which is much higher than the usual concentration of ATP in healthy tissues (Di Virgilio et al. 2018). 100-500µM ATP is enough to activate the purinergic receptor P2X7. To investigate the impact of P2X7 activation on the immune system in the tumor microenvironment, I injected 4T1 cells expressing P2X7(k) into Balb/C mice. When the tumor size reached 100mm3, the mice were randomized into four different groups and treated with intraperitoneal injections of PBS, an agonistic P2X7 nanobody (14D5.HLE), DOX or the combination of 14D5 and DOX. The tumor size was measured every second day. Interestingly, the tumors grew slower in the 14D5 treated group compared to placebo (Fig. 26A), supporting our previous observations that the
presence of P2X7 (k) on tumors is beneficial for the host (Section 3.2.1 and 3.2.2).
DOX alone did not show any decrease in tumor size while combination therapy (14D5+DOX) caused a slight decrease in tumor growth compared to placebo (Fig. 26B).
Figure 26. P2X7 enhancing nanobody slows down the tumor growth: 4T1/P2X7(k)
cells (5*105) were injected into the mammary fat pads of Balb/c mice. At day 8, when the tumor growth reached 100mm3, the mice were randomized into different groups and injected intraperitoneally twice a week with PBS, DOX (3mg/kg), 14D5 HLE (2mg/kg) or a combination of DOX and 14D5. Tumor volume was assessed in-vivo by caliper at the indicated time points. XY graph in the figure represents the increase in tumor volume vs days post 4T1 P2X7 (K) cells injection. A) 14D5 treated mice (red
curve) showed slower tumor growth compared to Placebo (black curve). B) DOX
treated mice did not show any reduction in tumor size. Statistics: Two-way analysis of variance; posttest: Dunnett's test. *, p<0.05.
6 8 10 12 15 17 21 24 26 29
0 500 1000 1500
Days post injection
Tu m o r v o lu m e ( m m 3 ) ø DOX 14D5 + DOX 6 8 10 12 15 17 21 24 26 29 0 500 1000 1500
Days post injection
Tu m o r v o lu m e ( m m 3 ) ø 14D5 * * * A B
